Skip to main content
Statistics LibreTexts

22.6: Bayes Factor

  • Page ID
    8837
  • We discussed Bayes factors in the earlier chapter on Bayesian statistics – you may remember that it represents the ratio of the likelihood of the data under each of the two hypotheses: K=P(data|HA)P(data|H0)=P(HA|data)*P(HA)P(H0|data)*P(H0) K = \frac{P(data|H_A)}{P(data|H_0)} = \frac{P(H_A|data)*P(H_A)}{P(H_0|data)*P(H_0)} We can compute the Bayes factor for the police search data using the contingencyTableBF() function from the BayesFactor package:

    ## Bayes factor analysis
    ## --------------
    ## [1] Non-indep. (a=1) : 1.8e+142 ±0%
    ## 
    ## Against denominator:
    ##   Null, independence, a = 1 
    ## ---
    ## Bayes factor type: BFcontingencyTable, independent multinomial

    This shows that the evidence in favor of a relationship between driver race and police searches in this dataset is exceedingly strong.